Well logging rock debris description method and device

Abstract

The invention provides a logging rock debris description method and device. The method comprises the following steps: acquiring a rock debris picture; segmenting the rock debris photo to enable each rock debris particle to have an independent area, and obtaining a segmented image; identifying the segmented image by using a rock debris particle identification model, and determining particle information in the rock debris photo according to an identification result; and performing rock debris description on the rock debris photo according to the particle information in the rock debris photo. By utilizing the scheme of the invention, the rock debris identification and description of the microscopic particle level can be realized, and the logging efficiency and accuracy are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of geological exploration technology, specifically to a method and apparatus for describing logging cuttings. Background Technology

[0002] Currently, the method for identifying logging cuttings involves workers collecting samples of cuttings, which are then visually assessed by geologists who provide corresponding descriptive information. The real-time nature and accuracy of cuttings identification are entirely controlled by the workers and geologists. Furthermore, with increasingly faster drilling speeds and finer cuttings particles, manual identification is no longer sufficient to meet the demands of modern logging operations. Summary of the Invention

[0003] This invention provides a method and apparatus for describing logging cuttings, so as to realize intelligent description of logging cuttings at the logging site.

[0004] Therefore, the present invention provides the following technical solution:

[0005] A method for describing logging cuttings, the method comprising:

[0006] Obtain photos of rock cuttings;

[0007] The rock debris photograph is segmented so that each rock debris particle has an independent region, resulting in a segmented image;

[0008] The segmented image is identified using a rock debris particle recognition model, and the particle information in the rock debris photograph is determined based on the recognition results.

[0009] Describe the rock cuttings based on the particle information in the rock cuttings photograph.

[0010] Optionally, the rock cuttings photographs include white light photographs and fluorescence photographs;

[0011] The particle information corresponding to the white light photograph includes any one or more of the following: color, mineral name, rounded or broken shape, cement, and particle characteristics;

[0012] The particle information corresponding to the fluorescence photograph includes: oiliness.

[0013] Optionally, segmenting the rock debris photograph so that each rock debris grain has an independent region includes:

[0014] The rock debris images were segmented using the watershed algorithm;

[0015] Region recognition is performed on the segmented image to classify each particle into a segmentation box;

[0016] If the segmentation box is a contiguous image, then continue segmenting the contiguous image until each particle has its own segmentation box.

[0017] Optionally, the method further includes:

[0018] Combine information on different types of rock cutting particles to establish a logical rule table for rock cutting description;

[0019] The step of describing the rock cuttings based on the particle information in the rock cuttings photograph includes:

[0020] Based on the particle information in the rock cuttings photograph, the rules in the rock cuttings description logic rule table are matched sequentially to obtain the matching result;

[0021] Describe the rock cuttings in the rock cuttings photographs based on the matching results.

[0022] Optionally, the rock fragment description includes any one or more of the following: color description, oil-bearing description, and lithological characteristic description.

[0023] Optionally, the method further includes:

[0024] Collect a large number of rock debris photographs;

[0025] Each particle in the rock debris photograph is labeled to generate training samples with labeled information;

[0026] A rock debris particle recognition model was trained using the training samples.

[0027] Optionally, the method further includes:

[0028] Save the training samples to the sample library;

[0029] Determine whether the recognition result is correct;

[0030] If the labeling information of the rock cuttings photograph is incorrect, then the labeling information of the rock cuttings photograph is corrected and simultaneously saved to the sample library.

[0031] A logging cuttings description device, the device comprising:

[0032] The photo acquisition module is used to acquire photos of rock cuttings;

[0033] The segmentation module is used to segment the rock debris photograph so that each rock debris particle has an independent region, resulting in a segmented image;

[0034] The identification module is used to identify the segmented image using a rock debris particle identification model, and to determine the particle information in the rock debris photograph based on the identification result;

[0035] The description module is used to describe the rock cuttings based on the particle information in the rock cuttings photograph.

[0036] Optionally, the segmentation module includes:

[0037] A segmentation unit is used to segment the rock debris photograph using the watershed algorithm;

[0038] The region recognition unit is used to perform region recognition on the segmented image and divide each particle into a segmentation box.

[0039] The judgment unit is used to determine whether the segmentation box is attached to the image. When the segmentation box is attached to the image, the segmentation unit is triggered to continue segmenting the attached image until each particle has its own segmentation box.

[0040] Optionally, the device further includes:

[0041] The rule table creation module is used to combine information on different types of rock cutting particles to create a logical rule table for describing rock cuttings.

[0042] The description module sequentially matches the rules in the rock cuttings description logic rule table with the particle information in the rock cuttings photo to obtain the matching result; and performs rock cuttings description on the rock cuttings photo based on the matching result.

[0043] Optionally, the device further includes: a model training module for training a rock cuttings particle recognition model; the model training module includes:

[0044] The photo acquisition unit is used to acquire a large number of rock debris photos;

[0045] The sample generation unit is used to label each particle in the rock debris photograph and generate training samples with labeling information.

[0046] The training unit is used to train a rock debris particle recognition model using the training samples.

[0047] Optionally, the device further includes:

[0048] A sample library is used to store the training samples;

[0049] The identification result detection module is used to determine whether the identification result is correct; if it is incorrect and the annotation information of the rock cuttings photo is wrong, the annotation information of the rock cuttings photo is corrected and saved to the sample library simultaneously.

[0050] A computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to perform the steps of the logging cuttings description method.

[0051] A computer program product includes a computer program / instructions that, when executed by a processor, implement the steps of the logging cuttings description method.

[0052] The logging cuttings description method and apparatus provided by this invention segment cuttings photographs, giving each cuttings particle an independent region. A cuttings particle recognition model is used to identify the segmented images, obtaining the cuttings identification results at the microscopic particle level, i.e., particle information. Then, cuttings description is performed on the cuttings photographs based on this particle information. Using this invention, on-site cuttings description can be digitized, unified, and standardized, achieving intelligent lithology description and significantly improving the efficiency of logging operations. Furthermore, this solution can reduce the workload on-site, providing technical support for future automated and unmanned logging.

[0053] Furthermore, combining information on different types of rock cuttings and establishing a logical rule table for rock cutting description can make the description of rock cuttings more standardized.

[0054] Furthermore, by collecting a large number of rock debris photos and annotating them at the micro-particle level, training samples with annotation information are generated. The rock debris particle recognition model is trained using the training samples, and this model can be used to achieve finer-grained rock debris recognition.

[0055] Furthermore, through manual evaluation of the rock cuttings identification results, errors can be detected, corrected, and added to the database in a timely manner. By continuously expanding the sample database, the rock cuttings identification model can be continuously optimized to achieve better rock cuttings identification results. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0057] Figure 1 This is a flowchart of a logging cuttings description method provided by the present invention;

[0058] Figure 2 This is a flowchart of a rock chip particle recognition model trained in an embodiment of the present invention;

[0059] Figure 3 This is a schematic diagram of rock cuttings photographs with annotation information in an embodiment of the present invention;

[0060] Figure 4 This is a schematic diagram of a logging cuttings description device provided in an embodiment of the present invention;

[0061] Figure 5This is another structural schematic diagram of the logging cuttings description device provided in an embodiment of the present invention;

[0062] Figure 6 This is a schematic diagram of a model training module in one embodiment of the present invention;

[0063] Figure 7 This is another structural schematic diagram of the logging cuttings description device provided in an embodiment of the present invention. Detailed Implementation

[0064] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0065] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0066] like Figure 1 The diagram shown is a flowchart of a logging cuttings description method provided by the present invention, which includes the following steps:

[0067] like Figure 1 The diagram shown is a flowchart of a logging cuttings description method provided by the present invention, which includes the following steps:

[0068] Step 101: Obtain photos of rock cuttings.

[0069] Specifically, photographic equipment can be used to acquire various rock debris images. To ensure image clarity and improve the accuracy of rock debris identification, certain camera parameters can be set. For example, in a non-limiting embodiment, the following camera parameters can be set: a 65-megapixel camera with 7.2x optical zoom, capable of clearly imaging rock debris particles with diameters ranging from 0.1mm to 4.0mm within a 60mm × 40mm field of view, and an image size of 9344 pixels × 7000 pixels.

[0070] The rock cuttings photographs may include: white light photographs and / or fluorescent photographs. Using white light photographs, the rock cuttings particle information that can be identified includes, but is not limited to, any one or more of the following: color, mineral name, rounded or broken shape, cement, particle characteristics, etc. Using fluorescent photographs, the rock cuttings particle information that can be identified mainly includes: oiliness.

[0071] Step 102: Segment the rock debris photograph so that each rock debris particle has an independent region, and obtain the segmented image.

[0072] Whether it's a white light photograph or a fluorescence photograph, the segmentation method is similar. For example, in a non-limiting embodiment, a watershed algorithm can be used to segment a rock debris photograph. The segmented image is then used for region identification, assigning each particle to a segmentation box (e.g., a rectangle). The segmentation boxes are checked to see if they are contiguous images; if so, the contiguous images are further segmented until each particle has its own segmentation box. In other words, each particle is separated from the overall photograph, resulting in multiple segmentation boxes, each containing one particle. Therefore, the region defined by each segmentation box corresponds to a segmented image.

[0073] A single photograph of rock debris can be segmented into multiple bounding boxes.

[0074] Step 103: Use the rock debris particle recognition model to identify the segmented image, and determine the particle information in the rock debris photo based on the recognition result.

[0075] The rock debris particle recognition model is a microscopic particle level rock debris recognition model that can automatically identify each segmented image, that is, automatically identify the particles within the segmentation box to obtain the particle information in the image.

[0076] Since the oil content of rock cuttings can be determined based on their color, separate models for identifying the lithology and oil content of rock cutting particles can be established. In other words, the rock cutting particle identification model can include both a lithology identification model and / or an oil content identification model. Using a white-light photograph of rock cuttings and the lithology identification model, the lithological information of the particles in the photograph can be obtained; using a fluorescent photograph of rock cuttings and the oil content identification model, the oil content of the particles in the photograph can be obtained.

[0077] Specifically, a single white-light photograph of a rock fragment can be segmented into multiple grain images. These grain images are then individually identified using a rock fragment grain lithology identification model, yielding the identification result for each grain. Combining these results allows us to obtain the lithological information of most grains in the rock fragment photograph. For example, the identified grain lithology information includes: quartz, gray, non-crystalline, and sub-angular. Furthermore, the proportion of each lithological feature can be calculated.

[0078] Similarly, using the fluorescence photograph of the rock fragment, multiple particle images can be obtained. The oil content identification model of the rock fragment particles can be used to identify these particle images one by one to obtain the oil content identification result for each particle image. By combining these identification results, the oil content information of most particles in the rock fragment photograph can be obtained, such as the oil content identification result being that it contains heavy oil.

[0079] The lithological and oil content identification results were combined to obtain the particle information of the rock fragment: heavy oil-quartz-gray-nothing-sub-angular.

[0080] Step 104: Describe the rock cuttings in the rock cuttings photograph based on the particle information in the rock cuttings photograph.

[0081] To ensure consistency in the description of rock cuttings, in a non-limiting embodiment, rock cutting particle information can be pre-combined to establish a logical rule table for rock cutting description. It should be noted that separate rule tables for describing rock composition and characteristics, and for describing rock cutting color and oil content, can be established.

[0082] For example, a table describing the composition and characteristics of rock fragments may include: a nameable component, corresponding rock fragment grains, rock fragment characteristics, characteristic type, description number, and descriptive statement. The nameable component may include, for example, sandstone, conglomerate, etc. The rock fragment grains may include information such as their size. The rock fragment characteristics may include, for example, rounded, angular, sub-rounded, sub-angular, etc. The characteristic type may include, for example, content, composition, cementing material, etc. The description number is a unique identifier for each type of rock fragment grain. The descriptive statement is a textual description of this type of rock fragment grain, such as: sandstone content xx, contains a small amount of XX, rounded, etc.

[0083] For example, the table describing the color and oil content of rock cuttings can also include: a name, and corresponding rock cutting particles, rock cutting characteristics, characteristic type, description number, and descriptive statement. The name can include, for example, oil-rich, oil stains, oil traces, fluorescence, no oil, and mixed colors. The rock cutting particles can include particle name and oil content. The rock cutting characteristics can include the color and oil content grade of the rock cuttings. The description number is a unique identifier for each type of rock cutting particle. The descriptive statement is a textual description of this type of rock cutting particle, such as: predominantly ##1, containing a small amount of ##3, and the named rock cuttings accounting for xx% of the rock cuttings.

[0084] Accordingly, after identifying the particle information in the rock cuttings photograph, the rules in the corresponding rock cuttings description logic rule table can be matched sequentially according to the particle information in the rock cuttings photograph to obtain the matching result; the rock cuttings photograph is described according to the matching result, thereby realizing the automatic identification and description of rock cuttings.

[0085] Furthermore, custom rule modules and formula editor modules can be developed based on the rock cuttings description rule table. The lithology of the rock cuttings can be automatically determined by combining the particle information such as composition, color, oil content, roundness, and cementation obtained through model identification with the calculated particle size and proportion. This can be done through logical combinations such as "AND", "OR", "NOT", and "None".

[0086] It should be noted that, in specific implementation, the content of the rock cuttings description can be determined according to the actual application needs. For example, it may include any one or more of the following: color description, oil-bearing description, and lithological characteristic description.

[0087] For describing mixed colors, first determine if there are at least three types of colors present (more than 10% of the particle color). Represent these colors according to their concentration as variables ##1, ##2, ##3, and ##4. If ##1 accounts for 50%–100%, ##2 for 25%–50%, ##3 for 10%–25%, and ##4 for 1%–10%, the color description is "predominantly ##1, secondarily ##2, with a small amount of ##3, and a trace of ##4." If ##1, ##2, and ##3 all account for 25%–50%, the color description is "predominantly ##1, ##2, and ##3." If ##1 and ##2 account for 25%–50%, and ##3 for 10%–25%, the color description is "predominantly ##1 and ##2, with a small amount of ##3."

[0088] For example, in describing oil content, the computer first calculates the percentage of oil-bearing and oil-containing rock fragments, and the percentage of named rock fragments. The ratio of these ratios represents the percentage of oil-bearing rock fragments in the named rock fragments. If the ratio ranges from 5% to 40%, it is classified as an oil-bearing spot. The description statement "Rock fragments in the named rock fragments percentage ##1, oil-bearing rock fragments in the named rock fragments percentage ##2, fluorescence color yellow (or dark yellow), occurrence porphyritic" is then invoked.

[0089] Taking gravelly, unevenly grained sandstone as an example, the description of rock fragment characteristics is as follows: First, calculate HL1 = (labeled "sandstone" + calculated "sandstone") / (total sand + gravel content) × 100%. Then, calculate the percentage of coarse sand, medium sand, fine sand, and silt, with the largest percentage being XX1, the second largest being XX2, and the third largest being XX3. The description would then be "Sandstone accounts for HL1, mainly XX1, followed by XX2, with a small amount of XX3." Descriptions such as rounded, sub-rounded, sub-angular, angular, sub-angular to sub-rounded are determined and described based on the percentage of grain roundness. HL2 = (labeled "conglomerate" + calculated "conglomerate") / (total sand + gravel content) × 100%. Obtain the major axis length LJ1 and minor axis length LJ2 of the largest grain, and the major axis length LJ3 and minor axis length LJ4 of the average grain. The description is: "Gravel accounts for HL2, the composition is mainly rock fragments and quartz, followed by feldspar, containing a small amount of dark minerals, the largest gravel diameter is LJ1×LJ2mm, the smallest is 1×1mm, generally LJ3×LJ4mm, poorly sorted." The determination of cement is also based on the proportion of argillaceous cement, siliceous cement, and cement in the identified particle information. For example, if the percentage of argillaceous cement is greater than 50%, it is described as "argillaceous cement, loose."

[0090] For example, in the table of rules for describing the color and oil content of rock cuttings, rules 1.6-1.7 describe the color as mixed. If the grain information in the rock cuttings photo is identified as having 50% gray, which meets rule 1.6.1, and dark gray accounts for 30% and red accounts for 15%, then the description statement "The color is mainly gray, followed by dark gray, with a small amount of red" will be output.

[0091] For example, in the particle information of the obtained rock debris photos, sub-rounded particles account for 30%, sub-angular particles account for 30%, rounded particles account for 10%, and angular particles account for 20%, which meets rule 1.9, and the output description statement is "sub-rounded ~ sub-angular".

[0092] For example, based on the particle information in the identified rock cuttings photos, the calculation (content of oil-bearing particles and oil-containing particles) ÷ sandstone yields an oil content of 50%, which conforms to rule 1.16, and the output of the oil-bearing description statement "Oil-bearing rock cuttings account for 50% of the named rock cuttings".

[0093] For example, based on the particle information in the identified rock fragment photos, the maximum gravel diameter of 5.1×7.5 mm is automatically calculated, which conforms to rule 1.7, and the description statement is output as "The composition is mainly rock fragments and quartz, followed by feldspar, with a small amount of dark minerals. The maximum gravel diameter is 5.1 mm × 7.5 mm, and the minimum gravel diameter is 1 mm × 1 mm".

[0094] The above description rules and matching methods are merely illustrative examples. The specific rules can be determined according to the requirements of the actual application, such as the granularity of differentiation. This embodiment of the invention does not limit these rules.

[0095] This invention also provides a method for constructing a rock fragment identification model. The rock fragment identification model is a rock fragment identification model at the micro-particle level. For the lithology and oil content of rock fragments, corresponding identification models can be constructed respectively, namely, a rock fragment lithology identification model and a rock fragment oil content identification model.

[0096] For the lithology identification model of rock fragments, a large number of white light photos of rock fragments can be collected and manually labeled. The white light photos with labeled information can be used as training samples to train the lithology identification model of rock fragments.

[0097] Similarly, for the oiliness identification model of rock fragments, a large number of fluorescent photographs of rock fragments can be collected and manually labeled. The fluorescent photographs with labeled information can be used as training samples to train the rock fragment oiliness identification model.

[0098] Since the two identification models with different characteristics are constructed in a similar way, the following description will use the rock debris particle identification model to uniformly represent the two identification models with different characteristics, and will no longer describe them separately.

[0099] Reference Figure 2 The flowchart illustrating the construction of a rock debris particle identification model in an embodiment of the present invention includes the following steps:

[0100] Step 201: Collect a large number of rock debris photographs.

[0101] Specifically, photographic equipment can be used to acquire various rock debris images. To ensure image clarity and improve the accuracy of the particle recognition model, certain camera parameters can be set. These camera parameters can be set according to actual application needs, and this embodiment of the invention does not impose specific limitations on them.

[0102] For example, a 65-megapixel camera is used to photograph dry rock cuttings. A blue rock cutting background is required, and particles with a diameter of 0.05-4.0 mm can be clearly imaged.

[0103] Step 202: Label each particle in the rock debris photograph to generate a training sample with labeled information.

[0104] With the help of some existing image annotation tools, each particle in the rock debris photo can be manually segmented and labeled. The segmentation box can be a rectangle, and the labeling information for each rectangle can include, but is not limited to, any one or more of the following: mineral name, composition, oil content, color, rounding (fracture morphology), cement, particle characteristics (information related to naming or description), etc.

[0105] like Figure 3 The image shown is an example of a photograph of rock debris with annotation information.

[0106] By manually labeling a large number of rock cuttings, a file containing the labeling information of the rock cuttings can be obtained, such as a JSON file.

[0107] For each rock debris photograph, a training sample with labeled information can be generated.

[0108] Step 203: Use the training samples to train a rock debris particle recognition model.

[0109] In this embodiment of the invention, a rock debris particle recognition model can be obtained by training training samples using a deep learning algorithm. The specific structure of the rock debris particle recognition model is not limited in this embodiment; any conventional training method adapted to the specific structure of the model can be used.

[0110] Using the rock debris particle recognition model provided in this embodiment of the invention, it is possible to locate and detect rock debris particles with a particle size in the range of 0.05-4.0mm in the entire photo, and obtain particle information.

[0111] In some embodiments, the training samples may also be saved to a sample library. Accordingly, in Figure 1 After the segmented image is identified using the rock debris particle recognition model in step 103, manual assistance can be used to determine whether the recognition result is correct. If it is incorrect and the annotation information of the rock debris photo is wrong, the annotation information of the rock debris photo can be corrected and simultaneously saved to the sample library. This can update and improve the sample quality of the sample library. In addition, for particles that cannot be identified, they can be manually labeled and added to the sample library at any time. Correspondingly, the current rock debris particle recognition model can be optimized through newly added and / or updated samples to further improve the accuracy of the rock debris particle recognition model.

[0112] The logging cuttings description method provided by this invention segments logging cuttings photographs, giving each cuttings particle an independent region. A cuttings particle recognition model is then used to identify the segmented images, obtaining the cuttings identification results at the microscopic particle level, i.e., particle information. The logging cuttings photographs are then described based on this particle information. Using this invention, the description of logging cuttings in the field can be digitized, unified, and standardized, achieving intelligent identification and description of cuttings, which can greatly improve the efficiency of logging field work. Furthermore, this solution can reduce the workload of logging field operations, providing technical support for future automated and unmanned logging.

[0113] The logging cuttings description method provided in this invention is based on the information of each cuttings particle in the cuttings photograph, refined to the particle level, combining the proportion and combination relationship of different information of cuttings particles, and establishing a cuttings description logic rule table based on cuttings description standards. This ensures that the description results of similar cuttings are unique, avoiding the multiple interpretations and human experience factors that often arise in traditional logging methods that rely on geologists' descriptions. Compared to traditional logging methods that rely solely on geologists' eyes and experience to estimate percentage content, this method not only improves accuracy but also significantly increases logging efficiency.

[0114] Accordingly, embodiments of the present invention also provide a logging cuttings description device, such as... Figure 4 The diagram shown is a structural schematic of the device.

[0115] The logging cuttings identification and positioning device 400 includes:

[0116] Photo acquisition module 401 is used to acquire photos of rock cuttings;

[0117] The segmentation module 402 is used to segment the rock debris photograph so that each rock debris particle has an independent region, thereby obtaining a segmented image;

[0118] The identification module 403 is used to identify the segmented image using a rock debris particle identification model, and to determine the particle information in the rock debris photograph based on the identification result;

[0119] The description module 404 is used to describe the rock cuttings based on the particle information in the rock cuttings photograph.

[0120] One specific structure of the segmentation module may include the following units:

[0121] A segmentation unit is used to segment the rock debris photograph using the watershed algorithm;

[0122] The region recognition unit is used to perform region recognition on the segmented image and divide each particle into a segmentation box.

[0123] The judgment unit is used to determine whether the segmentation box is attached to the image. When the segmentation box is attached to the image, the segmentation unit is triggered to continue segmenting the attached image until each particle has its own segmentation box.

[0124] like Figure 5 As shown, in a non-limiting embodiment, the logging cuttings identification device 400 may further include: a rule table establishment module 405, used to combine different types of cuttings particle information to establish a cuttings description logic rule table.

[0125] Accordingly, the description module 404 can sequentially match the rules in the rock cutting description logic rule table according to the particle information in the rock cutting photograph to obtain the matching result; and perform rock cutting description on the rock cutting photograph according to the matching result.

[0126] In this embodiment of the invention, the rock cuttings identification model can be established by a corresponding model training module. This model training module can be part of the logging rock cuttings identification device of the present invention, or it can be independent of the logging rock cuttings description device. This embodiment of the invention does not limit this.

[0127] The rock debris particle recognition model can employ a deep neural network model, and one specific structure of the model training module is as follows: Figure 6 As shown.

[0128] The model training module 600 includes the following units:

[0129] Photo acquisition unit 601 is used to acquire a large number of rock debris photos;

[0130] The sample generation unit 602 is used to label each particle in the rock debris photograph and generate training samples with labeling information.

[0131] Training unit 603 is used to train a rock debris particle recognition model using the training samples.

[0132] In order to continuously optimize the rock debris particle identification model, such as Figure 7 As shown, in another non-limiting embodiment of the logging cuttings description device of the present invention, it may further include: a sample library 406 and an identification result detection module 407. Wherein:

[0133] Sample library 406 is used to store the training samples;

[0134] The identification result detection module 407 is used to determine whether the identification result is correct; if it is incorrect and the annotation information of the rock cuttings photo is wrong, the annotation information of the rock cuttings photo is corrected and saved to the sample library simultaneously.

[0135] Accordingly, in this embodiment, the model training module 600 can optimize the rock debris particle recognition model using newly added and / or updated samples, such as by performing optimization training at certain time intervals or after a certain number of new samples have been accumulated. Through continuous model optimization, the accuracy of the rock debris particle recognition model can be further improved.

[0136] The specific implementation methods of the above modules can be referred to the description in the previous embodiments of the present invention, and will not be repeated here.

[0137] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0138] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0139] In the several embodiments provided by the present invention, it should be understood that the disclosed apparatus can be implemented in other ways.

[0140] The present invention also provides a storage medium, which is a computer-readable storage medium storing a computer program thereon, the computer program being executable when it runs. Figure 1 and Figure 2 The method shown may include some or all of the steps. The storage medium may include read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, etc. The storage medium may also include non-volatile memory or non-transitory memory, etc.

[0141] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means.

[0142] The embodiments of the present invention have been described in detail above. Specific implementation methods have been used to illustrate the present invention. The descriptions of the embodiments above are only for the purpose of helping to understand the methods and systems of the present invention, and are merely some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention, and the content of this specification should not be construed as a limitation of the present invention. Therefore, any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for describing logging cuttings, characterized in that, The method includes: Obtain photos of rock cuttings; The rock debris photograph is segmented so that each rock debris particle has an independent region, resulting in a segmented image; The segmented image is identified using a rock debris particle recognition model, and the particle information in the rock debris photograph is determined based on the recognition results. Describe the rock cuttings based on the particle information in the rock cuttings photograph.

2. The logging cuttings description method according to claim 1, characterized in that, The rock cuttings photographs include white light photographs and fluorescent photographs; The particle information corresponding to the white light photograph includes any one or more of the following: color, mineral name, rounded or broken shape, cement, and particle characteristics; The particle information corresponding to the fluorescence photograph includes: oiliness.

3. The logging cuttings description method according to claim 1, characterized in that, The segmentation of the rock debris photograph, so that each rock debris particle has an independent region, includes: The rock debris images were segmented using the watershed algorithm; Region recognition is performed on the segmented image to classify each particle into a segmentation box; If the segmentation box is a contiguous image, then continue segmenting the contiguous image until each particle has its own segmentation box.

4. The logging cuttings description method according to claim 1, characterized in that, The method further includes: Combine information on different types of rock cutting particles to establish a logical rule table for rock cutting description; The step of describing the rock cuttings based on the particle information in the rock cuttings photograph includes: Based on the particle information in the rock cuttings photograph, the rules in the rock cuttings description logic rule table are matched sequentially to obtain the matching result; Describe the rock cuttings in the rock cuttings photographs based on the matching results.

5. The logging cuttings description method according to claim 4, characterized in that, The description of the rock fragments includes any one or more of the following: color description, oil-bearing description, and lithological characteristic description.

6. The logging cuttings description method according to any one of claims 1 to 5, characterized in that, The method further includes: Collect a large number of rock debris photographs; Each particle in the rock debris photograph is labeled to generate training samples with labeled information; A rock debris particle recognition model was trained using the training samples.

7. The logging cuttings description method according to claim 6, characterized in that, The method further includes: Save the training samples to the sample library; Determine whether the recognition result is correct; If the labeling information of the rock cuttings photograph is incorrect, then the labeling information of the rock cuttings photograph is corrected and simultaneously saved to the sample library.

8. A logging cuttings description device, characterized in that, The device includes: The photo acquisition module is used to acquire photos of rock cuttings; The segmentation module is used to segment the rock debris photograph so that each rock debris particle has an independent region, resulting in a segmented image; The identification module is used to identify the segmented image using a rock debris particle identification model, and to determine the particle information in the rock debris photograph based on the identification result; The description module is used to describe the rock cuttings based on the particle information in the rock cuttings photograph.

9. The logging cuttings description device according to claim 8, characterized in that, The segmentation module includes: A segmentation unit is used to segment the rock debris photograph using the watershed algorithm; The region recognition unit is used to perform region recognition on the segmented image and divide each particle into a segmentation box. The determination unit is used to determine whether the segmentation box is attached to the image. When the segmentation box is attached to the image, the segmentation unit is triggered to continue segmenting the attached image until each particle has its own segmentation box.

10. The logging cuttings description device according to claim 8, characterized in that, The device further includes: The rule table creation module is used to combine information on different types of rock cutting particles to create a logical rule table for describing rock cuttings. The description module sequentially matches the rules in the rock cuttings description logic rule table with the particle information in the rock cuttings photo to obtain the matching result; and performs rock cuttings description on the rock cuttings photo based on the matching result.

11. The logging cuttings description device according to any one of claims 8 to 10, characterized in that, The device further includes: a model training module for training a rock debris particle recognition model; the model training module includes: The photo acquisition unit is used to acquire a large number of rock debris photos; The sample generation unit is used to label each particle in the rock debris photograph and generate training samples with labeling information. The training unit is used to train a rock debris particle recognition model using the training samples.

12. The logging cuttings description device according to claim 11, characterized in that, The device further includes: A sample library is used to store the training samples; The identification result detection module is used to determine whether the identification result is correct; if it is incorrect and the annotation information of the rock cuttings photo is wrong, the annotation information of the rock cuttings photo is corrected and saved to the sample library simultaneously.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run by the processor, it performs the steps of the logging cuttings description method according to any one of claims 1 to 7.

14. A computer program product comprising a computer program / instructions, characterized in that, When executed by a processor, the computer program / instruction implements the steps of the logging cuttings description method according to any one of claims 1 to 7.

Images